Endpoint failure: original admissible tests with uniform full-background reserve
ProvedConnesGreen.canonical_quartet_endpoint_failure_iff_uniform_full_background_testsFor each actual critical-strip zeta zero ρ off the critical line, there exists one positive attained quartet cut c ≥ positiveSupportRadius with the exact closed inner half-bound characterization: for every positive T, half ≤ G_Q(T) iff T ≤ c. At this same cut, FAILURE of the prescribed ordered support-right half-bound is equivalent to: there exists ONE real accuracy η with 0 < η < 1/2 such that for EVERY positive larger original window T > c there exists an ORIGINAL smooth compactly supported test g, with topological support inside (−T,T), satisfying the quantitative COMPLETE-background reserve
κ (Re W(gstarInv(g)) + ||B_Q,T x_g||²) < −(κ−1)||N_Q,T* x_g||², where κ=(1/2−η)⁻¹−1>1 and x_g is the original source embedding of Lg. The SAME test also satisfies ||B_Q,T* x_g||² < −Re W(gstarInv(g)) and Re W(gstarInv(g)) < 0. B_Q,T is the complete actual-negative complement outside the unchanged quartet, with ALL analytic multiplicities and paired/reflected normalizations preserved. No background zero is discarded.
The cut and η are chosen before T; g may depend on T. The first inequality gives a fixed relative coefficient, not one common test or a fixed absolute negative magnitude. The theorem asserts an exact equivalence, NOT endpoint failure or these test witnesses unconditionally. No finite-off-axis hypothesis, critical-endpoint identification, uniform arithmetic estimate, spectral inverse, prescribed Mellin values or RH assumption is introduced. The prescribed regularization then support-right limit order is preserved. RH remains open. Original RPB108 form attachments remain future formalization work.
import Definitions.Def_ConnesGreen_original_quartet import Definitions.Def_ConnesGreen_RG0_original_support_right_marker import Definitions.Def_WeilMarker_regularized_cost import Definitions.Def_ConnesGreen_small_support_constants open Complex MeasureTheory ConnesRZ ConnesRZFrontier ConnesRZQuartet ConnesGreen WeilDefect WeilDefect.ConnesNative WeilDefect.MarkerStability open scoped BigOperators InnerProductSpace lp ENNReal Classical ComplexOrder Topology open ContinuousLinearMap Filter Set set_option autoImplicit false set_option maxHeartbeats 3000000 set_option synthInstance.maxHeartbeats 1000000 set_option backward.isDefEq.respectTransparency false noncomputable section
theorem ConnesGreen.canonical_quartet_endpoint_failure_iff_uniform_full_background_tests
(ρ : CriticalZeros) (hoff : ρ.1.re ≠ 1 / 2) :
∃ c : ℝ, ∃ hc : 0 < c, positiveSupportRadius ≤ c ∧
(∀ T : ℝ, ∀ hT : 0 < T,
((1 / 2 : ℝ) • (1 : ℓ²({τ : CriticalZeros // τ ∈ quartet ρ}, ℂ) →L[ℂ]
ℓ²({τ : CriticalZeros // τ ∈ quartet ρ}, ℂ)) ≤ canonicalPicardMarker T hT (quartet ρ) ↔ T ≤ c)) ∧
(¬ (1 / 2 : ℝ) • (1 : ℓ²({τ : CriticalZeros // τ ∈ quartet ρ}, ℂ) →L[ℂ]
ℓ²({τ : CriticalZeros // τ ∈ quartet ρ}, ℂ)) ≤ canonicalSupportRightMarker c hc.le (quartet ρ) ↔
∃ η : ℝ, 0 < η ∧ η < 1 / 2 ∧
∀ T : ℝ, ∀ hT : 0 < T, c < T →
∃ g : ℝ → ℂ, SupportedTest T g ∧
(((1 / 2 - η)⁻¹ - 1) *
((weilDistribution (conv g (starInv g))).re +
‖(canonicalBackgroundSynthesis T hT (quartet ρ)).adjoint
(sourceEmbed T (problemOneL g))‖ ^ 2) <
-(((1 / 2 - η)⁻¹ - 1) - 1) *
‖(canonicalSelectedSynthesis T hT (quartet ρ)).adjoint
(sourceEmbed T (problemOneL g))‖ ^ 2) ∧
‖(canonicalBackgroundSynthesis T hT (quartet ρ)).adjoint
(sourceEmbed T (problemOneL g))‖ ^ 2 <
-(weilDistribution (conv g (starInv g))).re ∧
(weilDistribution (conv g (starInv g))).re < 0) := by sorry